Analysis and exploitation of multipath ghosts in radar target image classification.
other
Where this comes from
- Record sourced from PubMed, PMID 24577193.
- Also identified by DOI 10.1109/TIP.2014.2302683.
- No licence information is recorded for this record.
- Because redistribution is not established, this page shows the abstract only. Follow the links below for the full text.
Abstract
An analysis of the relationship between multipath ghosts and the direct target image for radar imaging is presented. A multipath point spread function (PSF) is defined that allows for specular reflections in the local environment and can allow the ghost images to be localized. Analysis of the multipath PSF shows that certain ghosts can only be focused for the far field synthetic aperture radar case and not the full array case. Importantly, the ghosts are shown to be equivalent to direct target images taken from different observation angles. This equivalence suggests that exploiting the ghosts would improve target classification performance, and this improvement is demonstrated using experimental data and a naïve Bayesian classifer. The maximum performance gain achieved is 32%.